About me

I am a postdoctoral researcher specializing in machine learning, Bayesian inference, and high-performance computing, applied to large-scale scientific datasets. My work sits at the intersection of ML and physics: I build neural network surrogate models, probabilistic inference pipelines, and GPU-accelerated software to extract insight from cosmological observations and simulations.

I am currently a postdoc in the Cosmological Physics and Advanced Computing (CPAC) group at Argonne National Laboratory, where I develop and deploy ML-powered models (in JAX and PyTorch) on multi-node, multi-GPU HPC systems. My current focus is building fast, differentiable surrogate models that replace expensive physics simulations, enabling large-scale Bayesian inference workflows that would otherwise be computationally prohibitive.

Before Argonne, I completed my PhD at the Laboratoire de Physique Subatomique et Cosmologie (Université Grenoble Alpes, France), where I built end-to-end data analysis pipelines for high-resolution telescope observations, notably achieving a ~1000× speedup through software reengineering and algorithmic optimization and leading developments of Monte-Carlo simulation frameworks to validate analysis pipelines.